Machine Learning for Time Series

About this book

The Python time-series ecosystem is large and hard to get a good grasp on, with many libraries and models. This book deepens your understanding of time series through an overview of the popular Python packages, and helps you build better predictive systems.

It re-introduces the basics of time series, then builds up traditional autoregressive models alongside modern non-parametric ones. Through practical examples and the theory behind them, you'll grow confident loading time-series data from any source, working with deep learning models such as recurrent neural networks and causal convolutional networks, and applying gradient boosting with feature engineering.

The book helps you match the right model to the right problem by explaining the theory behind several useful models, with real-world case studies covering weather, traffic, biking, and stock market data.

Highlights

• Explore modern machine learning methods, including online and deep learning algorithms
• Increase prediction accuracy by matching the right model to the right problem
• Learn from real-world case studies in operations, marketing, finance, and healthcare

What you'll learn

• Understand the main classes of time series and detect outliers and patterns
• Choose the right method to solve time-series problems
• Characterize seasonal and correlation patterns with autocorrelation and statistical techniques
• Visualize time-series data effectively
• Understand classical models like ARMA and ARIMA
• Implement deep learning models including Gaussian processes and transformers
• Work with libraries like Prophet, XGBoost, and TensorFlow

Who this book is for

For data analysts, data scientists, and Python developers who want practical recipes to use today and a reference for tomorrow. Basic knowledge of Python is required, and familiarity with statistics helps you get the most from the book.